APPLICATION OF NEURAL NETWORKS IN GIS

Abstract

A geographic information system is a very powerful tool for managing and analyzing land use data. The integration of geographic information systems and artificial neural networks offers a mechanism to reduce the cost of landscape change analysis by reducing the amount of time spent interpreting data. Artificial neural networks (ANNs) have been proven to be useful in interpreting natural resource information. Backpropagation neural networks are one of the most common and widely used architectures. Many ANN architectures and types have been developed, many of them PC-based. Prediction of changes is based on Markov chain analysis. This process determines the state of a system based on its previous state and the likelihood of changes occurring in between. Change models serve as useful tools for studying the different mechanisms by which land use changes occur, actual design and potential future environmental impacts, and impact assessments.